New course in FMI
by Vasil KolevWe’ll have a new course in FMI – different case studies of the system and network administration with Linux. The course is every Thursday in hall 325 of FMI from 19:00 (some days it might be from 18:00, we’ll notify the people beforehand. We have a site for the course, the current list of teachers is me, Boyan Krosnov, Marian Marinov and Nikolay Nedyalkov (I’ll also talk with Vlado Karavelov on the last cases with the hardware).
Here’s the conspectus:
00– architectures used in the case studies
01– recovery – building recovery systems
02– config management – svn/git
03– basic router – firewall, routing, openvpn, nat, bridge, bonding
04– policy routing/multihoming+NAT
N links, incoming DNAT, connmark
N links, outgoing SNAT, balancing
N links, load balancing/connection (2 subcases)
05– traffic control
latency optimization
fairness
deep packet inspection (layer 7 filtering)
06– dynamic routing
OSPF
BGP
VRRP
anycast
07– heartbeat failover
drbd/iscsi/FC
OCFS
application tracking (mysql)
08– load balancing
DNS loadbalancing
layer 4 loadbalancing (iptables/LVS)
layer 7 loadbalancing (squid/tomcat)
09– network troubleshooting
path mtu
layer 2 cases – unknown unicast flooding, mac address table problems
bgp problems – youtube problem, long prefix problem, uRPF
filter tracking (hping, tcptraceroute)
asymetric routing problems (looking glass)
10– system/kernel debugging
sigsegv tracing (gdb, ulimit), attaching gdb to live process
strace, ltrace (limitations) – ffmpeg case, libdb exim case
11– optimization
dynamic content/apache
static content
DB optimizations – indexes, query optimizations, explain
select * from pesho where pesho.a in (select gosho from table2);
hardware recommendations – db, router, storage, app server, voip server